U.S. patent application number 13/114261 was filed with the patent office on 2012-11-29 for method of normalizing a fluorescence analyzer.
This patent application is currently assigned to IDEXX LABORATORIES, INC.. Invention is credited to Robert W. LaChapelle, James Edward Milan, Garland Christian Misener.
Application Number | 20120300205 13/114261 |
Document ID | / |
Family ID | 47217688 |
Filed Date | 2012-11-29 |
United States Patent
Application |
20120300205 |
Kind Code |
A1 |
Misener; Garland Christian ;
et al. |
November 29, 2012 |
METHOD OF NORMALIZING A FLUORESCENCE ANALYZER
Abstract
A method of normalizing an analyzer response value of a
fluorescence analyzer is provided. The method includes measuring an
excitation spectrum of the analyzer and measuring an emission
sensitivity spectrum of the analyzer. Next, a normalization factor
based at least in part upon the excitation spectrum of the analyzer
and the emission sensitivity spectrum of the analyzer is
determined. The sample is then analyzed to obtain an uncorrected
analyzer response value. A normalized analyzer response value is
calculated based at least in part upon the uncorrected analyzer
response value and the normalization factor.
Inventors: |
Misener; Garland Christian;
(Portland, ME) ; Milan; James Edward; (Buxton,
ME) ; LaChapelle; Robert W.; (Leeds, ME) |
Assignee: |
IDEXX LABORATORIES, INC.
Westbrook
ME
|
Family ID: |
47217688 |
Appl. No.: |
13/114261 |
Filed: |
May 24, 2011 |
Current U.S.
Class: |
356/317 |
Current CPC
Class: |
G01J 3/28 20130101; G01J
3/4406 20130101; G01N 2201/124 20130101; G01N 21/64 20130101; G01N
2021/6417 20130101 |
Class at
Publication: |
356/317 |
International
Class: |
G01J 3/30 20060101
G01J003/30 |
Claims
1. A method of normalizing an analyzer response value of a
fluorescence analyzer, the method comprising the steps of:
measuring an excitation spectrum of the analyzer; measuring an
emission sensitivity spectrum of the analyzer; determining a
normalization factor based at least upon the excitation spectrum of
the analyzer and the emission sensitivity spectrum of the analyzer;
analyzing a sample to obtain an uncorrected analyzer response
value; and calculating a normalized analyzer response value based
at least upon the uncorrected analyzer response value and the
normalization factor.
2. The method according to claim 1, wherein the excitation spectrum
is measured using a spectrometer and wherein the emission
sensitivity spectrum is measured using a monochromator.
3. The method according to claim 1, wherein the normalization
factor is determined in accordance with a target normalization
factor.
4. The method according to claim 1, wherein the normalization
factor is determined in accordance with at least one wavelength
characteristic function of the excitation spectrum and at least one
wavelength characteristic function of the emission sensitivity
spectrum.
5. A method of normalizing an analyzer, the method comprising the
steps of: measuring an excitation spectrum of the analyzer;
determining at least one wavelength characteristic function of the
excitation spectrum of the analyzer; measuring an emission
sensitivity spectrum of the analyzer; determining at least one
wavelength characteristic function of the emission sensitivity
spectrum of the analyzer; and determining a normalization factor
based at least upon the at least one wavelength characteristic
function of the excitation spectrum and the at least one wavelength
characteristic function of the emission sensitivity spectrum,
wherein, the normalization factor is used to normalize an analyzer
response value of the analyzer.
6. The method according to claim 5, wherein the excitation spectrum
of the analyzer is measured to obtain a data set of relative
irradiance as a function of wavelength of the excitation
spectrum.
7. The method according to claim 5, wherein the emission
sensitivity spectrum of the analyzer is measured to obtain a data
set of relative irradiance sensitivity as a function of wavelength
of the emission sensitivity spectrum.
8. The method according to claim 5, wherein, the at least one
wavelength characteristic function of the excitation spectrum
includes at least one of: a wavelength centroid function, a
function representing an area under at least a portion of the
excitation spectrum, a function representing a wavelength at which
a relative irradiance exceeds a first predetermined value, a
function representing a wavelength at which the relative irradiance
falls below a second predetermined value, and a function
representing a relative irradiance value at a particular
wavelength.
9. The method according to claim 5, wherein, the at least one
wavelength characteristic function of the emission sensitivity
spectrum includes at least one of: a wavelength centroid function,
a function representing an area under at least a portion of the
emission sensitivity spectrum, a function representing a wavelength
at which a relative irradiance sensitivity exceeds a first
predetermined value, a function representing a wavelength at which
the relative irradiance sensitivity falls below a second
predetermined value, and a function representing a relative
irradiance sensitivity at a particular wavelength.
10. The method according to claim 5, wherein, the normalization
factor is determined in accordance with at least one of a plurality
of wavelength characteristic functions of the excitation spectrum
and at least one of a plurality of wavelength characteristic
functions of the emission sensitivity spectrum, the at least one
wavelength characteristic function of the excitation spectrum and
the at least one wavelength characteristic function of the emission
sensitivity spectrum selected by comparison to a target
normalization factor.
11. The method according to claim 5, wherein the excitation
spectrum is measured using a spectrometer and wherein the emission
sensitivity spectrum is measured using a monochromator.
12. The method according to claim 5, wherein a set of normalization
factors is determined, each normalization factor corresponding to a
different sensor to be detected by the analyzer.
13. The method according to claim 12, wherein the normalization
factor for each sensor is determined in accordance with at least
one fitting parameter specific to the sensor.
14. A method of normalizing an analyzer in accordance with a
specific sensor to be detected by the analyzer, the method
comprising the steps of: measuring an excitation spectrum of the
analyzer; measuring an excitation spectrum of the sensor;
determining a first parameter based at least upon the overlap of
the excitation spectrum of the analyzer and the excitation spectrum
of the sensor; measuring an emission sensitivity spectrum of the
analyzer; measuring an emission spectrum of the sensor; determining
a second parameter based at least upon an overlap of the emission
sensitivity spectrum of the analyzer and the emission spectrum of
the sensor; and determining a normalization factor based at least
upon at least one of the first and second parameters, wherein, the
normalization factor is used to normalize an analyzer response
value of the analyzer and the sensor.
15. The method according to claim 14, wherein the step of measuring
the excitation spectrum of the sensor includes: measuring an early
excitation spectrum of sensor; measuring a late excitation spectrum
of the sensor; and wherein, the first parameter is a ratio of the
overlap of the excitation spectrum of the analyzer with the late
excitation spectrum of the sensor divided by the overlap of the
excitation spectrum of the analyzer with the early excitation
spectrum of the sensor.
16. The method according to claim 14, wherein the step of measuring
the emission spectrum of the sensor includes: measuring an early
emission spectrum of sensor; measuring a late emission spectrum of
the sensor; and wherein, the second parameter is a ratio of the
overlap of the emission sensitivity spectrum of the analyzer with
the late emission spectrum of the sensor divided by the overlap of
the emission sensitivity spectrum of the analyzer with the early
emission spectrum of the sensor.
17. The method according to claim 14, wherein at least one of the
excitation spectrum of the analyzer and the emission sensitivity
spectrum of the analyzer is measured to obtain a data set of
relative irradiance as a function of wavelength of the excitation
spectrum of the analyzer and the emission sensitivity spectrum of
the analyzer, respectively.
18. The method according to claim 14, wherein at least one of the
excitation spectrum of the sensor and the emission spectrum of the
sensor is measured to obtain a data set of relative fluorescence as
a function of wavelength of the excitation spectrum of the sensor
and the emission spectrum of the sensor, respectively.
19. The method according to claim 14, wherein the normalization
factor is determined in accordance with at least one fitting
parameter specific to the sensor.
20. The method according to claim 14, wherein the normalization
factor is determined in accordance with a target normalization
factor.
Description
BACKGROUND
[0001] 1. Technical Field
[0002] The present disclosure relates to fluorescence analysis and,
more particularly, to methods of normalizing fluorescence analyzers
according to specific wavelength characteristics thereof.
[0003] 2. Background of Related Art
[0004] Filter fluorometry refers generally to the process of
illuminating a sample and quantitating the resulting, detected
fluorescence signal. A filter fluorometer, designed to perform
fluorometry, employs one or more light sources, optical filters,
and detectors to select the wavelength ranges of both the
illuminating light and the light detected. Generally, these two
kinds of wavelength ranges do not overlap to any appreciable
extent, and the detection wavelengths are longer than the
illumination wavelengths. Using a predetermined calibration curve,
the detected signal is correlated to the amount of fluorescent
component present in the sample. For example, filter fluorometry
(hereafter, simply "fluorometry") may be used to measure the
concentration of certain electrolytes in blood plasma or serum by
their effect on the fluorescence of a test device, e.g., a
slide.
[0005] Currently, electrolyte detection via fluorometry involves
the use of sensors which include reagents, e.g.,
electrolyte-specific fluoroionophores (for Na.sup.+ and K.sup.+
detection) or acridinium reagents (for Cl.sup.- detection).
Initially, a slide containing only the appropriate sensor (e.g.,
the "dry" slide) is read by an analyzer designed to perform
fluorometry on such slides, in order to measure an "early"
fluorescence intensity. This "early" fluorescence may be an average
"early" fluorescence intensity of multiple "early" fluorescence
intensity measurements. Next, the sample to be tested is added to
and allowed to interact with the slide. The slide, including the
sample (e.g., the "wet" slide), is again read by the analyzer in
order to measure a "late" fluorescence intensity. Similarly as
above, this "late" fluorescence intensity may be an average of
multiple "late" fluorescence intensity measurements. These measured
values are then corrected in accordance with a fluorescence
baseline and an analyzer response is calculated based on the ratio
of the baseline-corrected "early" and "late" fluorescence intensity
measurements. More specifically, the analyzer response is
calculated according to:
AR e = I late , e - I baseline , e I early , e - I baseline , e EQ
1 ##EQU00001##
Where "AR" is the analyzer response for the given electrolyte, "e,"
and wherein the "l" values are the measured fluorescence
intensities (or averages thereof). Note that "early" fluorescence
intensity measurement may also be obtained soon after sample
addition.
[0006] The analyzer response "AR," may then be used, in accordance
with the slide's lot calibration, to determine the specific
electrolyte's concentration in the sample under test. However,
there are multiple sources of variability, e.g.,
analyzer-to-analyzer variability (wavelength variation in the
fluorescence detection module, dispensed volume variation,
incubation temperature variation, timing variation, etc.) and also
sensor-to-sensor variability, that may produce different analyzer
responses for the same sample. It is therefore necessary to
correct, or "normalize" these analyzer responses in order to
accurately determine the electrolyte concentration in a given
sample.
[0007] Current methods of normalizing analyzer response values "AR"
typically require running "known" samples on each analyzer to be
normalized. More specifically, in one particular method of
normalizing analyzer responses, a target analyzer response value
for a given electrolyte, slide lot, and sample electrolyte
concentration is first assigned. Then, using that given slide lot,
the samples, which have different, but known electrolyte
concentrations, are run on the analyzer to be normalized, in order
to determine an uncorrected analyzer response for each sample.
Multiple runs may be performed, indexed, and averaged with one
another to determine an average uncorrected analyzer response for
each sample. These uncorrected analyzer responses are then compared
to the assigned target analyzer response values that correspond to
the given electrolyte concentrations in the samples to determine a
specific normalization factor, "s.sub.e,c," for each of the
samples, e.g., for each given electrolyte concentration. This
normalization factor, "s.sub.e,c," is used for all slide lots,
although it is only calculated based upon a given slide lot, e.g.,
the given slide lot mentioned above. Next, the electrolyte
concentration-specific normalization factors are averaged (or
otherwise correlated) to yield an electrolyte-specific
normalization factor applicable to all concentrations of that
electrolyte. This normalization factor, "s.sub.e," in turn, is used
to correct, or "normalize" the analyzer response, e.g., according
to:
AR e , corr = AR e s e EQ 2 ##EQU00002##
[0008] This corrected, or "normalized" analyzer response,
"AR.sub.e,corr," in conjunction with the slide's lot calibration,
is then used to determine the electrolyte concentration of a
specific electrolyte in an "unknown" sample.
[0009] Although the above-described normalization process, and
similar test-based normalization processes, correct for a number of
sources of variability, e.g., dispensed volume variation,
incubation temperature variation, and timing variation, such
processes also introduce a number of additional sources of
variability, including: accurate determination of the target
analyzer response values, accuracy of the sample concentrations,
and the applicability to all slide lots of a single normalization
factor determined for one slide lot. Further, the above-described
methods also require that multiple samples be run on each analyzer
to be normalized in order to determine the normalization factor
used to normalize the analyzer response Therefore, a need exists to
develop a more efficient and more accurate method for normalizing
analyzer response values in fluorometry.
[0010] It has been found that a major source of variability in
fluorometry for electrolyte detection results from variation in
wavelength characteristics of the filter fluorometer. Further, it
has been found that accounting for these wavelength-based
variations alone (and not considering the other sources of
variation discussed above: dispensed volume variation, incubation
temperature variation, and timing variation) yields better overall
precision, e.g., less analyzer-to-analyzer variation, as compared
to slide-based normalization methods, such as those described
above. However, producing filters and source(s) (e.g., LED(s)) for
use in the analyzers that are optimized to the required individual
analyzer accuracy such that wavelength variation is reduced to
insignificant levels would significantly increase the costs
associated with the manufacture and qualification of such
analyzers. It would therefore be desirable to provide a method of
wavelength-based normalization for reducing analyzer-to-analyzer
variation in fluorometry.
SUMMARY
[0011] A method of normalizing an analyzer response value of a
filter fluorescence analyzer is provided in accordance with one
embodiment of the present disclosure. The method includes measuring
an excitation irradiance vs. wavelength spectrum (hereafter, simply
"excitation spectrum") of the analyzer as well as sample emission
sensitivity vs. wavelength spectrum (hereafter, simply "emission
sensitivity spectrum") of the analyzer. A normalization factor is
determined based at least in part upon the excitation spectrum of
the analyzer and the emission sensitivity spectrum of the analyzer.
A sample is then analyzed, i.e., run through the analyzer, to
obtain an uncorrected analyzer response value. A normalized
analyzer response value can then be determined based at least in
part upon the uncorrected analyzer response value and the
normalization factor.
[0012] In one embodiment, the analyzer excitation spectrum is
measured using a spectrometer, while the analyzer emission
sensitivity spectrum is measured using a monochromator.
[0013] In another embodiment, the normalization factor is
determined in accordance with a target normalization factor.
Further, the normalization factor may be determined in accordance
with one or more wavelength characteristic functions of the
analyzer excitation spectrum and/or one or more wavelength
characteristic functions of the analyzer emission sensitivity
spectrum.
[0014] In accordance with another embodiment of the present
disclosure, a method of normalizing an analyzer is provided. The
method includes measuring an excitation spectrum of the analyzer
and determining one or more wavelength characteristic functions of
the excitation spectrum of the analyzer. The method further
includes measuring an emission sensitivity spectrum of the analyzer
and, similarly, determining one or more wavelength characteristic
functions of the emission sensitivity spectrum of the analyzer. A
normalization factor is then determined based at least in part upon
the wavelength characteristic function(s) of the excitation
spectrum and the wavelength characteristic function(s) of the
emission sensitivity spectrum. The normalization factor, as can be
appreciated, is used to normalize an analyzer response value of the
analyzer.
[0015] In one embodiment, the excitation spectrum of the analyzer
is measured to obtain a data set of relative (scaled) irradiance as
a function of wavelength. Similarly, the emission sensitivity
spectrum of the analyzer may be measured to obtain a data set of
analyzer signal to scaled (relative) input irradiance as a function
of wavelength.
[0016] In another embodiment, the wavelength characteristic
function(s) of the analyzer excitation spectrum include a
wavelength centroid function, a function representing an area under
a portion of the excitation spectrum, the relative irradiance at a
predetermined wavelength, a function representing a wavelength at
which a relative irradiance exceeds a first predetermined value,
and a function representing a wavelength at which the relative
irradiance falls below a second predetermined value. The wavelength
characteristic function(s) of the analyzer emission sensitivity
spectrum may include similar functions.
[0017] In yet another embodiment, the normalization factor is
determined in accordance with one (or more) of a plurality of
wavelength characteristic functions of the analyzer excitation
spectrum and/or one (or more) of a plurality of wavelength
characteristic functions of the analyzer emission sensitivity
spectrum. The one (or more) functions are selected by comparison to
a target normalization factor.
[0018] Similarly as in the previous embodiment, the analyzer
excitation spectrum may be measured using a spectrometer, while the
analyzer emission sensitivity spectrum may be measured using a
monochromator.
[0019] In still yet another embodiment, a set of normalization
factors is determined. In such an embodiment, each normalization
factor corresponds to a different sensor, e.g., the active
component of a slide, to be detected by the analyzer. Further, the
normalization factor for each sensor is determined in accordance
with one (or more) fitting parameter(s) specific to that
sensor.
[0020] In accordance with still another embodiment of the present
disclosure, a method of normalizing an analyzer in accordance with
a specific sensor to be detected by the analyzer is provided. The
method includes measuring excitation and emission sensitivity
spectra of the analyzer and excitation and emission spectra of the
sensor. A first parameter based at least in part upon the overlap
of the excitation spectrum of the analyzer and the excitation
spectrum of the sensor is determined. A second parameter based at
least in part upon an overlap of the emission sensitivity spectrum
of the analyzer and the emission spectrum of the sensor is
determined. A normalization factor is then determined based at
least in part upon the first parameter and/or the second parameter.
The normalization factor, as in the previous embodiments, is used
to normalize an analyzer response value of the analyzer and the
sensor.
[0021] In one embodiment, both an early excitation spectrum of the
sensor and a late excitation spectrum of the sensor are measured.
Similarly, both an early emission spectrum of the sensor and a late
emission spectrum of the sensor may be measured. In such an
embodiment, the first parameter is a ratio of the overlap of the
excitation spectrum of the analyzer with the late excitation
spectrum of the sensor divided by the overlap of the excitation
spectrum of the analyzer with the early excitation spectrum of the
sensor.
[0022] In another embodiment, the second parameter is a ratio of
the overlap of the emission sensitivity spectrum of the analyzer
with the late emission spectrum of the sensor divided by the
overlap of the emission sensitivity spectrum of the analyzer with
the early emission spectrum of the sensor.
[0023] In still another embodiment, the excitation spectrum of the
sensor and/or the emission spectrum of the sensor are measured to
obtain a data set of fluorescence intensity as a function of
excitation wavelength and/or fluorescence intensity as a function
of emission wavelength, respectively.
[0024] In yet still another embodiment, one or more fitting
parameters specific to the interaction of the sensor spectra with
the analyzer spectra are used in calculating the normalization
factor of that sensor. Further, a target normalization factor may
be used in calculating the normalization factor of a particular
sensor.
BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Various embodiments of the present disclosure are described
herein with reference to the drawings, wherein:
[0026] FIG. 1 is a schematic illustration of a configuration for
calibrating a monochromator for module emission sensitivity
spectrum measurement;
[0027] FIG. 2 is a schematic illustration of a configuration for
calibrating a monochromator for slide excitation spectrum
measurement;
[0028] FIG. 3 is a schematic illustration of a configuration for
measuring an excitation spectrum of a module under test;
[0029] FIG. 4 is an example module excitation spectrum;
[0030] FIG. 5 is a schematic illustration of a configuration for
measuring an emission sensitivity spectrum of the module;
[0031] FIG. 6 is an example module emission sensitivity
spectrum;
[0032] FIG. 7 is a schematic illustration of a configuration for
measuring a slide excitation spectrum;
[0033] FIG. 8 shows superimposed "early" and "late" excitation
spectra for sodium detection slides;
[0034] FIG. 9 is a schematic illustration of a configuration for
measuring the slide emission spectrum; and
[0035] FIG. 10 shows superimposed "early" and "late" emission
spectra for sodium detection slides.
DETAILED DESCRIPTION
[0036] Chemical analyzers, such as those disclosed in
commonly-owned, co-pending U.S. Patent Application Publication No.
2010/0254854, the entire contents of which are hereby incorporated
by reference herein, measure and analyze optical characteristics of
a sample to determine the presence and/or concentration of specific
molecules in the sample. In one particular example, chemical
analyzers are used to illuminate a sample in order to measure the
fluorescence intensity of the sample to determine the concentration
of specific molecules, e.g., electrolytes, in the sample. This
process is generally referred to as fluorometry.
[0037] Although the present disclosure is described with specific
reference to electrolyte detection, it is envisioned that the
present disclosure may similarly be applied to fluorometry in
general. In the following description, the terms "module" and
"analyzer" are generally meant to both describe a fluorescence
detection system, e.g., the "analyzer" referring to the system as a
whole and the "module" referring to the fluorescence detection
portion thereof, and, thus, are used generally interchangeably
hereinbelow. Further, for purposes of brevity, the term "slide" is
meant to include both the slide itself and the sensor, fluorophore,
or other electrolyte detector contained thereon. The slide's lot
calibration represents a specific calibration factor or function
specific to that set of slides. The slide's lot calibration may be
provided by the manufacturer of the slides, e.g., in the form of a
look-up table, or may be independently calculated. However, the
specific calibration of such is not central to the present
disclosure and, thus, will not be described herein.
[0038] As will be described in greater detail below, the present
disclosure relates to wavelength-based methods of normalizing
analyzer response values, e.g., the fluorescence intensity readings
from the fluorescence detection module of the analyzer, in order to
provide more accurate electrolyte concentration readings and to
reduce variability in electrolyte concentration readings from
analyzer-to-analyzer. These normalized fluorescence readings, in
accordance with the slide's lot calibration, are then used to
determine the concentration of a specific electrolyte within the
sample. As can be appreciated, the presently disclosed
normalization methods help reduce the effects of bias that a
specific module (or group of modules) of an analyzer may have on
the measured electrolyte concentrations of samples run using that
specific module (or group of modules). As such, the presently
disclosed wavelength-based normalization methods increase the
precision and uniformity of electrolyte concentration measurements
across all normalized analyzers. More specifically, the presently
disclosed wavelength-based normalization methods use wavelength
characteristics of the fluorescence detection module of the
analyzer and, in some embodiments, of the electrolyte-specific
sensors, e.g., fluorophores corresponding to the particular
electrolyte to be detected, in order to determine a normalization
factor specific to that analyzer for a given electrolyte to be
detected. The analyzer response reading of any sample run using the
analyzer may then be used, in accordance with the slide's lot
calibration information and the corresponding normalization factor,
to more accurately determine the concentration of a specific
electrolyte within the sample.
[0039] In one embodiment, as will be described in greater detail
below, one or more functions of the fluorescence detection module's
excitation spectrum characteristics and one or more functions of
the module's emission sensitivity spectrum characteristics are used
to determine the normalization factor for that particular module.
More specifically, in equation form, the electrolyte-specific
normalization factor, s.sub.e, for a particular analyzer is
determined in accordance with the function(s) of excitation
spectrum characteristics, f(.lamda.).sub.e,ex, and emission
sensitivity spectrum characteristics, f(.lamda.).sub.e,em, of the
module, generally according to:
s.sub.e=c.sub.e,em.times.f(.lamda.).sub.e,em+c.sub.e,ex.times.f(.lamda.)-
.sub.e,ex+c.sub.e,0 EQ 3:
where c.sub.e,em, c.sub.e,ex, and c.sub.e,0 are
electrolyte-specific fitting parameters for the module emission
sensitivity function, f(.lamda.).sub.e,em, the module excitation
function, f(.lamda.).sub.e,ex, and an offset, respectively.
[0040] In another embodiment, as will also be described in greater
detail below, the fluorescence detection module's excitation
spectrum is overlapped with the excitation spectrum of the specific
sensor, e.g., fluorophore, to be detected, and the module's
emission sensitivity spectrum is likewise overlapped with the
emission spectrum of the sensor, in order to determine the
normalization factor of the module. This relationship, or overlap,
indicates any overlap bias of a particular module/electrolyte
combination due to the fact that wavelength characteristics of
different modules overlap differently with different fluorophores
to be detected. Mathematically, the electrolyte-specific
normalization factor, s.sub.e, is determined in accordance with
overlapping function(s) of excitation spectrum characteristics of
the module, f(.lamda..sub.i).sub.m,ex, and the early and late
excitation spectrum characteristics of the sensor, or fluorophore,
to be detected, f(.lamda..sub.i).sub.e,early,ex,
f(.lamda..sub.i).sub.e,late,ex, respectively, and the overlapping
function(s) of emission sensitivity spectrum characteristics of the
module, f(.lamda.).sub.m,em, and the early and late emission
spectrum characteristics of the fluorophore to be detected,
f(.lamda..sub.i).sub.e,early,em, f(.lamda..sub.i).sub.e,late,em,
respectively, generally according to:
s e = c e , 1 .times. ( i = 1 N ex f ( .lamda. i ) m , ex .times. f
( .lamda. i ) e , late , ex i = 1 N ex f ( .lamda. i ) m , ex
.times. f ( .lamda. i ) e , early , ex ) .times. ( i = 1 N em f (
.lamda. i ) m , em .times. f ( .lamda. i ) e , late , em i = 1 N em
f ( .lamda. i ) m , em .times. f ( .lamda. i ) e , early , em ) + c
e , 0 EQ 4 ##EQU00003##
Where, similar to above, c.sub.e,1, and c.sub.e,0 are
electrolyte-specific fitting parameters and offsets, respectively.
Alternatively, as in the form of equation "EQ 3," above, the
excitation and emission spectra overlap ratios may each include a
separate fitting parameter, c.sub.e,ex and c.sub.e,em,
respectively, so that a weighted sum of the two ratios replaces the
product of these two ratios shown above.
[0041] The above overview of two exemplary methods for
wavelength-based normalization of electrolyte-detection modules is
not limited to those equations above. Rather, the equations above
are simply exemplary equations used to illustrate an example of how
wavelength characteristic functions of the module and/or sensor may
be used in determining the normalization factor for a particular
module/sensor pair. Each of the two embodiments discussed above
will be described in turn in greater detail below.
Calibration
[0042] As can be appreciated in view of the forgoing, in order to
employ wavelength-based normalization for a particular analyzer
(or, more particularly, for the fluorescence detection module
thereof) and electrolyte to be detected (or, more particularly,
sensor, or fluorophore, for detecting the particular electrolyte),
it is necessary to measure certain wavelength characteristics
thereof. These wavelength characteristics may include: the
excitation (output irradiance) spectrum of the module, the emission
sensitivity spectrum of the module, the excitation spectrum of the
slide, and/or the emission spectrum of the slide. More
specifically, in the first embodiment discussed above, the
normalization factors are determined based upon a function (or
functions) of the excitation spectrum of the module as well as a
function (or functions) of the emission sensitivity spectrum of the
module and, thus, both the excitation spectrum and the emission
sensitivity spectrum measurements are required. In the second
embodiment discussed above, the normalization factors are
determined based upon the overlap between the excitation spectrum
of the module and the excitation spectrum of the slide and the
overlap between the emission sensitivity spectrum of the module and
the emission spectrum of the slide, and, thus, these measurements
are required. The configurations and processes for measuring each
of these wavelength characteristics will be described in greater
detail below.
[0043] With reference to FIGS. 1 and 2, and as can be appreciated,
the accuracy of any spectrum measurement depends, at least in part,
on the accuracy of the monochromator 10 and/or spectrometer 20 used
to make the measurements. Accordingly, the monochromator 10 and
spectrometer 20 must be calibrated prior to obtaining the above
measurements to ensure that the spectral characteristics of the
module 30 and/or the slide 40 can be accurately measured. More
specifically, the monochromator 10 is used in measuring the
emission sensitivity spectrum 34 (FIG. 6) of the module 30 and the
excitation spectrum 42 (FIG. 8) of the slide 40 and is thus
calibrated for each of such measurements. The spectrometer 20, on
the other hand, is used for measuring the excitation spectrum 32
(FIG. 4) of the module 30 and the emission spectrum 44 (FIG. 10) of
the slide 40, as well as for calibrating the monochromator 10
(FIGS. 1 and 2).
[0044] Initially, the spectrometer 20 is calibrated. More
specifically, the spectrometer 20 is calibrated for wavelength
accuracy, linearity of response, and irradiance sensitivity. The
spectrometer 20 used for module excitation and slide emission
spectrum measurements may be the Ocean Optics HR2000+ Spectrometer
and associated SpectraSuite Spectrometer Operating Software, both
available from Ocean Optics, Inc. of Dunedin, Fla., USA, although
other suitable spectrometers may also be used. The aforementioned
SpectraSuite Software provides procedures for and facilitates
calibration of the spectrometer 20 for wavelength accuracy,
linearity of response, and irradiance sensitivity. The wavelength
accuracy calibration process synchronizes the wavelength positions
reported by the spectrometer 20 with well-characterized mercury and
argon emission lines. Linearity calibration, on the other hand, is
performed to help ensure that, for any given wavelength, the
response will scale linearly, e.g., such that a doubling of the
irradiance will double the response (signal intensity) produced by
the spectrometer 20. Similarly, the irradiance sensitivity
calibration is performed to help ensure that the response (signal
intensity) produced by the spectrometer 20 at one wavelength can be
scaled to the response (signal intensity) produced by the
spectrometer 20 at another wavelength, e.g., such that, where the
irradiance at a first wavelength and the irradiance at a second
wavelength are equal, the signal intensity at both wavelengths is
equal.
[0045] With the spectrometer 20 calibrated as discussed above, the
monochromator 10 may then be calibrated in accordance with the
previously-calibrated spectrometer 20. The monochromator 10 used
for module emission sensitivity and slide excitation spectrum
measurements may be the Ocean Optics MonoScan 2000. While this
monochromator is calibrated by its manufacturer, "dependent"
calibration of the monochromator 10 has been found to assure a more
traceable, and more accurate, calibration of the system, e.g., the
monochromator 10 and the spectrometer 20, as a whole, and, thus,
produces more accurate measurements as compared to an "independent"
calibration of the monochromator 10. As can be appreciated, the
calibration of the monochromator 10, which will be described below,
provides the data necessary to correlate the step positions of the
monochromator 10 to its output dispersion band centroid wavelengths
and also allows the detector counts of the module 30 (see FIG. 5)
or photomultiplier tube 100 (see FIG. 7) to be scaled, or adjusted,
in accordance with the relative irradiance provided by the
monochromator 10 and light source 60 at that specific wavelength
for which the detector counts were obtained. The monochromator 10
is calibrated first for module emission sensitivity (FIG. 1), in
order to obtain module emission sensitivity spectrum measurements,
and then for slide excitation (FIG. 2), in order to obtain slide
excitation spectrum measurements (although the order may be
reversed).
[0046] Referring now to FIG. 1, the configuration for calibrating
the monochromator 10 for module emission sensitivity spectrum
measurements is shown. As shown, the 0.3 mm fiber 50 is coupled
between the output 62 of the broadband light source 60 and the
input 12 of the monochromator 10 to provide unfiltered light to the
monochromator 10. The output 14 of the monochromator 10, in turn,
is coupled to the input 22 of the spectrometer 20 via the emission
test 0.4 mm fiber 70 such that the spectrometer 20 can read the
output dispersion bands of the monochromator 10. The 0.3 mm fiber
50 and the emission test 0.4 mm fiber 70 are the same as those to
be used during the module emission sensitivity spectrum
measurements, discussed below, such that any bias introduced by the
fibers 50, 70 is substantially eliminated.
[0047] With continued reference to FIG. 1, during calibration, the
spectrometer 20 reads the output dispersion bands of the
monochromator 10, e.g., the spectrometer 20 provides relative
irradiance values as a function of wavelength for each incremented
step position of the monochromator 10. Using this data, as will be
described below, monochromator step positions are
wavelength-calibrated in, e.g., nanometer (nm) units. Further, the
spectrometer data is also used to scale, or adjust the detector
counts produced by the module 30 at a given monochromator
wavelength to account for the relative output irradiance produced
by the monochromator 10 and light source 60 at that wavelength. In
other words, this data provides the information necessary to
correlate the module 30 detector count values of the analyzer into
relative irradiance sensitivity values.
[0048] In order to calibrate the monochromator 10, the dispersion
band spectrum produced by the monochromator 10 is obtained using
the spectrometer 20. More particularly, the monochromator 10 is
driven, in approximately 10 nm intervals (although other intervals
are contemplated), substantially along the wavelength range to be
calibrated and the spectrometer 20 is used to obtain the dispersion
band spectrum data at each interval therefrom. Each band spectrum
is represented as relative irradiance vs. wavelength (in nm)
ordered pairs.
[0049] Next, each dispersion band spectrum is least-squares fitted
to an un-normalized Gaussian function to determine a band centroid,
.lamda..sub.0, a band maximum, l.sub.0, and a bandwidth parameter,
.sigma., for each band. The monochromator step positions used are,
in turn, then least-squares fitted to the corresponding
Gaussian-fitted band centroids, .lamda..sub.0, determined above, by
a cubic equation.
[0050] Although the dispersion bands obtained in this way are
substantially Gaussian, they are not perfectly Gaussian and, thus,
correction is required. More specifically, it has been found that
the band maximum value estimates, l.sub.0, are too high. To
compensate for this, all data points within +/-1 nm of the band
centroids, .lamda..sub.0, for each dispersion band spectrum are
averaged; the averages of these values are then reassigned as the
band maximum values, l.sub.0, for the particular dispersion band
spectrum.
[0051] Next, these monochromator dispersion band maxima, l.sub.0,
are least-squares fitted to the previously determined band
wavelength centroids, .lamda..sub.0, by another cubic equation, to
determine the relative irradiance output of the monochromator 10
and source 60 as a function of wavelength.
[0052] Finally, an output correction function is determined for
converting the detector count values to relative irradiance values.
The output correction function may be as follows:
g ( .lamda. ) = .alpha. 0 + .alpha. 1 .lamda. low + .alpha. 2
.lamda. low 2 + .alpha. 3 .lamda. low 3 .alpha. 0 + .alpha. 1
.lamda. + .alpha. 2 .lamda. 2 + .alpha. 3 .lamda. 3 EQ 5
##EQU00004##
where .lamda..sub.low is the lowest wavelength in the monochromator
calibration and the .alpha. values are the cubic equation fitting
parameters. Using this calibration function, g(.lamda.), the module
30 detector count values can be scaled to relative irradiance
sensitivity values, e.g., via multiplying by g(.lamda.), for module
emission sensitivity spectrum measurements.
[0053] However, even if the monochromator 10 is accurately
calibrated, e.g., as described above, it remains important to
ensure that, upon start-up, the monochromator 10 is consistently
driven to a set "home" position. As can be appreciated, if the
monochromator 10 is not consistently set to the same "home"
position each time the monochromator 10 is initialized, the
wavelength positions obtained by driving the monochromator's
wavelength dispersion grating will vary. In other words, while the
monochromator's step size is accounted for in the above-described
calibration, if the monochromator 10 does not start from the same
"home" position after each start-up, the wavelength values
associated with the steps will be shifted by the difference between
the actual starting position and the "home" position.
[0054] In order to correct for such variability, a "soft home"
process is implemented during start-up of the monochromator 10.
More specifically, at each start-up, the monochromator grating (not
explicitly shown) is driven out of its initial position by a
pre-determined number of steps and is then returned to the initial
position. This procedure helps ensure that the initial, or start-up
position of the monochromator 10 corresponds to a consistent "home"
position. Thereafter, the monochromator 10 may be accurately driven
to the desired wavelength position.
[0055] Two other monochromator wavelength inaccuracy mitigations
are also implemented to ensure accurate calibration of the
monochromator 10. First, motor backlash is eliminated by driving
the monochromator 10 to a position before its starting wavelength,
i.e., to a lower value than the starting wavelength, prior to
driving the monochromator 10 to the starting wavelength. The
wavelength may then be scanned upward to reach the desired starting
wavelength. Second, the monochromator calibration process described
above is repeated multiple times, e.g., three times. The results
from each calibration are then combined to determine the wavelength
and irradiance sensitivity calibrations.
[0056] Turning now to FIG. 2, the configuration for calibrating the
monochromator 10 for slide excitation spectrum measurements is
shown. The configuration for calibration of the monochromator 10
for slide excitation spectrum measurements is similar to the
configuration for calibrating the monochromator 10 for module
emission sensitivity spectrum measurements (see FIG. 1). However,
rather than using the 0.3 mm fiber 50 to couple the light source 60
and monochromator 10 and the 0.4 mm emission test fiber 70 to
couple the monochromator 10 and the spectrometer 20, the 0.6 mm
fiber 80 is coupled between the output 62 of the light source 60
and the input 12 of the monochromator 10 to provide unfiltered
light to the monochromator 10, while the output 14 of the
monochromator 10 is coupled to the input 22 of the spectrometer 20
via the excitation test 1.0 mm fiber 90. Similarly as described
above, these fibers 80, 90 are the same as those to be used during
the slide excitation spectrum measurements such that any bias due
to the fibers 80, 90 is substantially eliminated. These larger core
diameter fibers 80, 90 are used for slide fluorescence excitation
spectrum measurements to ensure that the fluorescence signal is
strong enough for reproducible measurement.
[0057] The data produced by the spectrometer 20 is then used,
similarly as above, to correlate the step positions of the
monochromator 10 to its output wavelength centroids (e.g.,
nanometers) and to scale the photomultiplier tube 100 (see FIG. 7)
detector count values according to the relative input irradiance
from the monochromator 10 and source 60. The source 60 (and its
intensity setting) used for the monochromator 10 calibrations is
also used in obtaining module emission sensitivity spectra or slide
excitation spectra.
Measurement
[0058] FIGS. 3-10 show the configurations for measuring the various
wavelength characteristics of the module 30 and slide 40 and sample
spectra 32, 34 and 42, 44, respectively, for each of such
measurements. Each of these configurations will be discussed in
greater detail below. However, it is envisioned that other suitable
configurations may also be employed, as the configurations shown in
FIGS. 3, 5, 7 and 9 are only exemplary. Likewise, the sample
spectra shown in FIGS. 4, 6, 8 and 10 are just that, samples of
typical measured spectra. Further, the monochromator calibration
configurations discussed above (FIGS. 1-2) may also be modified, as
desired, and are similarly only exemplary configurations.
[0059] As shown in FIG. 3, the configuration for measuring the
excitation spectrum of a given module 30 under test is shown. As
shown, a 0.4 mm excitation test fiber 70 is coupled between the
module 30 and a spectrometer 20. For obtaining the module
excitation spectrum, the excitation test fiber 70 is approximately
centered on the optical axis of the module 30. This ensures that
the wavelengths received by the excitation test fiber 70 provide a
consistent module-to-module approximation of the wavelengths that
the slide 40 would receive during testing. Because the fluorescence
excitation intensity of the module 30 is high, the proximal end of
fiber 70 may be spaced back from the typical location at which
slide 40 is read. Further, the proximal end of fiber 70 is located
perpendicularly to the optical axis of module 30. The measured
excitation spectrum 32 (FIG. 4) of the module 30 provides the
relative excitation irradiance at each wavelength. More
specifically, this excitation spectrum 32 (FIG. 4) is used to
compile data in the form of a (wavelength, relative irradiance)
ordered pair for each wavelength in the range tested, typically in
0.2 nm increments. Further, if necessary, this data may then be
filtered, e.g., via Gaussian smoothing, boxcar averaging, or any
other suitable method. As can be appreciated, and as will be
described in greater detail below, this data indicates the presence
of any biases in the module 30 so that these biases may be
accounted for in the determination of the normalization factor
specific to that module 30.
[0060] With reference to FIG. 4, in conjunction with FIG. 3, a
sample module excitation spectrum 32 is shown. This module
excitation spectrum 32, as shown in FIG. 4, reveals a blue
wavelength band with a peak typically in the range of about 460 nm
to about 470 nm. The LED source (not explicitly shown) of the
module 30 and the filters (not explicitly shown) employed in the
module 30 work to limit the range of available excitation intensity
to within about 430 nm to about 500 nm. The module excitation
spectrum 32 shows the light energy provided by the module at a
range of wavelengths. More specifically, the module excitation
spectrum 32 represents the energy available to be absorbed by the
slide 40 during the fluorescence detection process for measuring
the concentrations of electrolytes in a sample. FIG. 4 also shows
two shaded areas under the spectrum curve corresponding to two
exemplary pre-determined wavelength ranges, 452 nm to 500 nm and
460 nm to 500 nm. As will be described in greater detail below, in
this example, these areas are ultimately used in order to estimate
the normalization factor to be employed for sodium slide tests.
[0061] With reference to FIG. 5, measurement of the emission
sensitivity spectrum 34 (FIG. 6) of the module 30 will be
described. Prior to measuring the emission sensitivity spectrum 34
(FIG. 6) of the module 30, the monochromator 10 is calibrated for
such measurement, as described above. The same 0.3 mm fiber 50 used
in the calibration of the monochromator 10 remains coupled between
the broadband light source 60 and the monochromator 10 for
providing unfiltered light to the monochromator 10. During each
test run, the monochromator 10, in turn, produces a narrow band of
light that is transmitted through a 0.4 mm emission test fiber 70
(the same fiber 70 used in the calibration of the monochromator 10)
coupled between the monochromator 10 and the module 30. The
proximal end of emission test fiber 70 is approximately centered on
and perpendicular to the optical axis of module 30 for the testing.
Further, the proximal end of the emission test fiber 70 is
positioned at approximately the same distance from module 30 as
slide 40 is to be located when the module 30 is used to read the
fluorescence of the slide 40. This positioning helps to ensure that
the wavelengths received from the emission test fiber 70 are
substantially similar to those that are to be received from the
slide 40. Typically, the tests are run at 0.2 nm increments across
the entire range of wavelengths to be tested. The resulting
emission sensitivity spectrum 34 (FIG. 6) as produced by the module
30 is then obtained in the form of (monochromator steps, detector
counts) ordered pairs. This data is subsequently scaled to the
corresponding (wavelength, relative input irradiance sensitivity)
ordered pairs. Similarly as above, the emission sensitivity
spectrum data may also be filtered. Further, as can be appreciated,
and as will be described in greater detail below, this data
indicates the presence of any biases in the module 30 so that these
biases may be accounted for in the determination of the
normalization factor specific to that module 30.
[0062] With reference to FIG. 6, a sample module emission
sensitivity spectrum 34 is shown. The uncorrected module emission
sensitivity spectrum 34 reveals a broad wavelength band of
sensitivity spanning a substantial portion of the green, yellow,
orange, and red regions of the electromagnetic spectrum. The
silicon photodiode detector (not explicitly shown) of the module 30
and the filters (not explicitly shown) employed in the module work
to limit the range of light sensitivity to within about 510 nm to
about 650 nm. The module emission sensitivity spectrum shows the
light energy sensitivity of the module at a range of wavelengths.
That is, the module emission sensitivity spectrum represents the
emitted slide energy that is detectable by the module during the
fluorescence detection process for measuring the concentrations of
electrolytes in a sample. FIG. 6 also shows the result of the
overlap of this module emission sensitivity spectrum (the
"uncorrected spectrum") with a composite (of the sodium, potassium,
and chloride electrolytes') dry slide emission spectra to form a
"corrected" spectrum. The corrected emission sensitivity spectrum
provides decreased sensitivity to variability towards the higher
end of the wavelength spectrum, i.e., it is less sensitive with
respect to longer wavelengths, as compared to the uncorrected
spectrum. By using the corrected module emission sensitivity
spectrum, more accurate module normalization factors can be
obtained.
[0063] FIG. 6 also shows two shaded areas under the corrected
emission sensitivity spectrum curve corresponding to two exemplary
pre-determined wavelength ranges, 500 nm to 544 nm and 500 nm to
556 nm. These areas, as will be described in greater detail below,
are ultimately used in to estimate the normalization factor to be
employed for sodium slide tests.
[0064] With reference now to FIG. 7, the excitation spectrum 42
(see FIG. 8) of the slide 40 may be overlapped with the excitation
spectrum 32 (see FIG. 4) of the module 30 in order to correct for
module-to-module and/or slide lot-to-slide lot wavelength
differences with specific normalization factors. As shown in FIG.
7, to obtain a fluorescence excitation spectrum of a sample, such
as of slide 40, the fluorescence emitted by the sample is detected
at a fixed wavelength band while scanning the input (excitation)
light wavelengths and recording the detected fluorescence intensity
for each wavelength scanned. As can be appreciated, generating a
narrow excitation wavelength band of light and transporting it to
the slide 40 is inefficient in that only a small fraction of the
source intensity is available to excite the fluorescence of the
slide 40. Therefore, a sensitive photomultiplier tube 100 is
employed to detect the slide fluorescence and is filtered, e.g.,
using a 550AELP redundant "long pass" filter 102, available from
Omega Optical, Inc. of Brattleboro, Vt., USA, to pass only light
wavelengths above about 550 nm. Similarly, large core diameter
fibers 90 and 110 are used to maximize light throughput. Like the
module emission sensitivity measurement (see FIG. 6), the slide
excitation spectrum 42 is essentially a sensitivity spectrum, in
this case detecting the fluorescence intensity in response to
excitation light of known relative (scaled) irradiance and
wavelength. The excitation light is delivered at an approximately
45.degree. angle to the slide 40, e.g., via the 1 mm excitation
test fiber 90, and the emitted light is detected in general axially
alignment with the center of the slide 40, e.g., via 1 mm detection
fiber 110.
[0065] With reference to FIG. 8, an example of sodium sensor
"early" and "late" spectra are shown. In order to obtain the
"early," or "dry," readings of slides 40 run in the analyzer,
readings are obtained at approximately 25 seconds and again at 9
seconds prior to the sample being added. These readings are then
averaged to obtain the "dry," or "early" slide reading. At
approximately 3 minutes and 14 seconds and again at 3 minutes and
25 seconds after the sample is added to the slide 40, the "wet"
slides 40 are once again read by the analyzer and a pair of "late"
readings for each slide 40 is obtained. These readings are then
averaged to form the "wet," or "late" slide reading. However, it is
also envisioned that different time intervals and/or greater or
fewer discrete readings may be analyzed, depending on the
particular sensor used. For example, the time intervals and the
number of readings used for sodium are different from those used
for potassium and chloride.
[0066] The "early" and "late" slide excitation spectra are obtained
at approximately the same time intervals as the "early" and "late"
slide readings, discussed above, in order to determine a
relationship between the spectra, which interact with the module
excitation spectrum, and the module's slide readings, e.g., in
order to determine the effect of the spectra on the module's slide
readings. This is important because, just as the sample
concentration affects the analyzer response, as discussed above
with respect to equation "EQ 1," the sample concentration also
affects the magnitude of the "late" spectrum at a given time after
sample addition, thereby affecting the analyzer response.
[0067] Further, as also discussed above with respect to equation
"EQ 1," above, the analyzer subtracts the background fluorescence
intensity, e.g., the baseline reading, from the early and late
readings prior to ratioing them to form the uncorrected analyzer
response. Similarly here, the fluorescent background baseline slide
spectra are obtained and are subtracted from the early and late
spectra. In obtaining these baseline spectra, special sensor slides
configured with the same ingredients and production processes as
the normal slides, minus the active fluorophore, are used. This
same type of "background" sensor slide is used in the analyzer,
typically at manufacture or service, to measure background
fluorescence intensity for subtraction from the early and late
fluorescence readings.
[0068] With reference to FIGS. 9 and 10, as mentioned above, the
emission spectrum 44 of the slide 40 may also be overlapped with
the emission sensitivity spectrum 34 of the module 30 (see FIG. 6)
to correct for module-to-module and/or slide lot-to-slide lot
wavelength differences with specific normalization factors. A
fluorescence emission spectrum 44 of a sample, such as of slide 40,
consists of recording the fluorescence emitted by the sample at
discrete wavelength values over a selected wavelength range, while
the input (excitation) light wavelength range and intensity remain
fixed. Due to the relatively weak fluorescence intensity and
insensitivity of the spectrometer, large core diameter fibers 64
and 90 are used to excite and capture sufficient fluorescence
intensity, respectively. A higher intensity source 61, e.g., a
Fiber-Lite.RTM. High Intensity Fiber Optic Illuminator, available
from Dolan-Jenner Industries of Boxborough, Mass., USA, and a short
pass (red blocking) filter 66, e.g., a 3RD490SP filter (Omega
Optical), provide the fluorescence excitation intensity required
for the slide emission spectrum measurement. Again, the excitation
light is delivered at an approximately 45.degree. angle to the
slide 40, and the emitted light is detected in general axial
alignment with the center of the slide 40.
[0069] With reference now to FIG. 10, examples of sodium sensor
"early" and "late" emission spectra are given. The spectra shown
are the results of obtaining "early" and "late" emission spectra at
approximately the same times as they are obtained by the analyzer,
and subtracting from each the sodium sensor background emission
slide spectrum, similarly as discussed above with reference to FIG.
8. As shown in FIG. 10, the two spectra do not have proportionate
relative fluorescence to one another; if they did, and the "early"
and "late" slide excitation spectra did as well, then the
uncorrected analyzer response ratio in equation "EQ 1" would
largely correct for analyzer-to-analyzer differences, and a
normalization factor would be unnecessary. However, as clearly
shown in FIG. 10, the shapes and peak wavelengths of the "early"
and "late" bands differ, thus requiring normalization. More
specifically, since different module emission sensitivity spectra
indicate variable relative sensitivity to the "early" and "late"
fluorescence spectra of a given sensor, thus resulting in different
analyzer responses, a normalization factor to correct for
module-to-module differences is required. Likewise, if the emission
spectra of the slide 40 vary from slide lot-to-slide lot, the
normalization factor may be calculated to adjust for these
differences as well. However, if the "early" and "late" emission
spectra remain consistent in form, other than multiplicative
scaling differences, which are compensated by slide lot
calibration, then the normalization factors need only be applied on
a module-to-module basis. Further, it is noted that, although
fluorescence intensities below 500 nm and above 660 nm are
produced, the module 30 has no sensitivity to fluorescence
intensity at wavelengths outside this about 510 nm to about 650 nm
range (see FIG. 6) and, thus, the slide emission spectra are only
measured within the range shown in FIG. 10.
[0070] The slide emission spectrum measurement configuration, e.g.,
the high-intensity broadband light source 61 (with short pass
filter 66) and excitation fiber 64, is calibrated using stable NIST
Standard Reference Material 2941 (not shown). The reference
material (not shown) is placed with its frosted slide located where
the slide 40 would normally be placed. The emission spectrum of the
reference material is obtained and compared against the
NIST-published reference emission spectrum for this standard.
Individual wavelength correction factors are determined by
computing the ratio of the NIST reference spectrum divided by the
measured emission spectrum of the standard for each wavelength
obtained and to be used for slide spectra. Slide emission spectra
are then corrected by multiplying the correction factor ratios by
the measured emission spectrum on a wavelength-by-wavelength
basis.
Target Normalization Factors
[0071] In some embodiments, target normalization factors may be
determined and are used in accordance with the measured module
wavelength characteristics to ultimately determine the
normalization factors of each analyzer for a given electrolyte to
be detected. Two different methods of determining the target
normalization factors for the module will be discussed in turn
below, although other suitable methods for determining target
normalization factors are also contemplated. Alternatively, it is
envisioned that the herein disclosed wavelength-based normalization
factors be calculated without reference to a target normalization
factor.
[0072] In one embodiment, the target normalization factors are
determined in accordance with a "standard analyzer" (or group of
analyzers). In this embodiment, a plurality of samples having
different concentrations is tested. More specifically, for each
sample run on the module under test, the electrolyte concentration
of the sample is also detected using the standard analyzer. The
slide's lot calibration information, and the electrolyte
concentration of the sample, as determined by the standard
analyzer, are then used to back-correlate a target analyzer
response, AR.sub.0,e,l,c, for each sample. The target normalization
factor, s.sub.e,c, for each concentration, in turn, may then be
calculated using the target analyzer response, AR.sub.0,e,l,c, and
the uncorrected analyzer responses, AR.sub.e,l,c,i (the analyzer
responses from running the sample on the module under test)
according to:
s e , c = i = 1 n AR e , l , c , i n .times. AR 0 , e , l , c EQ 6
##EQU00005##
As noted previously, e indicates the electrolyte to be detected, l
the slide lot, c the electrolyte sample concentration tested, and n
the total number of replicate tests, with i indexing the
replicates. Typically, these concentration-specific normalization
factors, s.sub.e,c, are averaged to determine the
electrolyte-specific (but not concentration-specific) target
normalization factor, s.sub.e, for the analyzer under test.
[0073] In another embodiment, a given sample (or samples) is run on
a "training set" of different analyzers, with each analyzer
producing an uncorrected analyzer response, AR.sub.e,l,c,i. These
uncorrected analyzer responses are then averaged to produce the
target analyzer response, AR.sub.0,e,l,c. From there, similarly as
above, an electrolyte-specific (but not concentration-specific)
target normalization factor, s.sub.e, for each analyzer is
determined.
[0074] Armed with the wavelength characteristic measurements and
target normalization factors, as determined by either of the
methods described above, or any other suitable method, the
normalization factors for a given module/fluorophore pair can be
determined, as will be described with reference to the embodiments
discussed below.
Embodiment 1
[0075] Initially, as discussed above, the module excitation
spectrum data, relative irradiance as a function of wavelength,
f(.lamda..sub.i).sub.m,ex,in, is collected as (wavelength, relative
irradiance) ordered pairs, and is indexed accordingly. The module
emission sensitivity data, detector counts as a function of
monochromator step, is converted to (wavelength, relative
irradiance sensitivity) ordered pairs in accordance with the
monochromator calibration information discussed above, to yield a
relative irradiance sensitivity spectrum as a function of
wavelength, f(.lamda..sub.i).sub.m,em,corr. This emission
sensitivity spectrum may be overlapped with a composite dry slide
emission spectrum to minimize the effects of portions of the module
emission sensitivity spectrum unimportant for calculating module
normalization factors. Further, as mentioned above, both data sets
may be filtered. The subscripts m, ex or em, and in or corr
indicate, respectively, the specific module, excitation or emission
sensitivity, and "input" or "corrected," respectively.
[0076] Next, the module excitation spectrum data and module
emission sensitivity data is standardized, to substantially remove
any random relative intensity calibration error. For example, the
module excitation spectrum data may be standardized according
to:
f ( .lamda. i ) m , ex = f ( .lamda. i ) m , ex , in i = 1 N ex f (
.lamda. i ) m , ex , in EQ 7 ##EQU00006##
while the module emission sensitivity spectrum data may be
standardized according to:
f ( .lamda. i ) m , em = f ( .lamda. i ) m , em , corr i = 1 N em f
( .lamda. i ) m , em , corr EQ 8 ##EQU00007##
[0077] Alternatively, it may be preferable to mathematically
normalize either one of the functions f(.lamda..sub.i).sub.m,ex,in
or f(.lamda..sub.i).sub.m,em,corr if the wavelength increments over
which the spectra were obtained are not equally-spaced. In either
case, the resulting functions f(.lamda..sub.i).sub.m,ex or
f(.lamda..sub.i).sub.m,em may be used to compare, respectively, the
excitation or emission sensitivity spectra of different
modules.
[0078] As another alternative, the spectra may be standardized
based upon the overlap of the respective excitation spectrum and
emission sensitivity spectrum of the module and a standard
(typically dry) slide excitation spectrum and emission spectrum. In
this embodiment, the standardized spectra are electrolyte-specific.
More particularly, the module excitation spectrum data may be
standardized according to:
f ( .lamda. i ) m , e , ex = f ( .lamda. i ) m , ex , in .times. f
( .lamda. i ) dry , e , ex i = 1 N ex f ( .lamda. i ) m , ex , in
.times. f ( .lamda. i ) dry , e , ex EQ 9 ##EQU00008##
while the module emission sensitivity spectrum data may be
standardized according to:
f ( .lamda. i ) m , e , em = f ( .lamda. i ) m , em , corr .times.
f ( .lamda. i ) dry , e , em i = 1 N ex f ( .lamda. i ) m , em ,
corr .times. f ( .lamda. i ) dry , e , em EQ 10 ##EQU00009##
These equations ("EQ 9" and "EQ 10") combine the spectrum overlap
process with a standardization process, both described above. A
preferred method for generating a "corrected,"
non-electrolyte-specific, composite dry slide emission spectrum's
overlap with a module emission sensitivity spectrum was described
above.
[0079] Using these standardized spectra, f(.lamda..sub.i).sub.m,ex
and f(.lamda..sub.i).sub.m,em, (or, for the electrolyte-specific
standardization, f(.lamda..sub.i).sub.m,e,ex and
f(.lamda..sub.i).sub.m,e,em) determined above, various
characteristics of the spectra can be determined, e.g., the
wavelength centroid, the area under portions of the spectrum, the
wavelength at which the relative irradiance exceeds or drops below
a particular level (or levels), the relative irradiance(s) at a
particular wavelength (or wavelengths). These characteristics are
then incorporated into one or more functions of such
characteristics that are then used in determining the corresponding
normalization factor.
[0080] For example, by testing various normalization factor
equation forms and least-squares optimizing functions of module
excitation and emission sensitivity spectra, the following
functions and coefficients were determined to be particularly
suitable for use in calculating the target normalization factors
for a given module:
s.sub.Na,m=a.sub.Na,0+a.sub.Na,ex(1)f(.lamda.).sub.m,ex,460+a.sub.Na,em(-
1)f(.lamda.).sub.m,em,556+a.sub.Na,pf(.lamda.).sub.m,ex,452f(.lamda.).sub.-
m,em,544 EQ 11:
s.sub.K,m=a.sub.K,0+a.sub.K,ex(1)f(.lamda.).sub.m,ex,458+a.sub.K,em(1)f(-
.lamda.).sub.m,em,556+a.sub.K,pf(.lamda.).sub.m,ex,452f(.lamda.).sub.m,em,-
544 EQ 12:
s.sub.Cl,m=a.sub.Cl,0+a.sub.Cl,emf(.lamda.).sub.m,em,536+a.sub.Cl,ex,1f(-
.lamda.).sub.m,ex,470+a.sub.Cl,ex,2f(.lamda.).sub.m,ex,470.sup.2 EQ
13:
[0081] In the electrolyte-specific, module-specific normalization
factor equations above, all f(.lamda.).sub.m values are fractional
areas under the curves of the standardized or mathematically
normalized excitation or emission sensitivity spectra for the
particular module tested. For example, f(.lamda.).sub.m,em,556
indicates the fractional area under the standardized or normalized
module emission spectrum curve from 500 nm through 556 nm.
Similarly, f(.lamda.).sub.m,ex,470 indicates the fractional area
under the standardized or normalized module excitation spectrum
from 470 nm through 500 nm. Note that 500 nm is here always the
starting or ending wavelength for module emission sensitivity or
excitation spectra, respectively. However, depending on a
particular purpose, the specific wavelength range (or ranges) used
to calculate these areas may be varied.
[0082] For "EQ 11" through "EQ 13" above, exemplary wavelength
function coefficients were determined by least-squares optimization
of data from a large training set of analyzers. In tabular form,
these coefficients are:
TABLE-US-00001 TABLE 1 Na and K coefficients a.sub.0 a.sub.ex(1)
a.sub.em(1) a.sub.p for Na norm. 0.668830 0.493135 0.946124
-1.592978 factor for K norm. 0.318983 0.588290 1.695024 -2.068804
factor
TABLE-US-00002 TABLE 2 Cl coefficients a.sub.0 a.sub.em a.sub.ex,1
a.sub.ex,2 for Cl norm. 0.935330 -0.066119 0.425202 -1.086057
factor
[0083] In general, the function, functions, or combinations thereof
that best fit the target normalization factors for a wide range of
analyzers are selected. Alternatively, the selected function,
functions, or combinations thereof may be used to assign a target
normalization factor for a module whose target normalization factor
is not known. In the first case, the normalization factor, s.sub.e,
is calculated in accordance with these functions of wavelength
characteristics, f(.lamda.).sub.e,ex and f(.lamda.).sub.e,em, to
optimize the fit between the calculated and target normalization
factor according to a function, as mentioned above, such as:
s.sub.e=c.sub.e,em.times.f(.lamda.).sub.e,em+c.sub.e,ex.times.f(.lamda.)-
.sub.e,ex+c.sub.e,0 EQ 14:
where c.sub.e,em, c.sub.e,ex, and c.sub.e,0 are optimization
parameters, where this equation is the same as "EQ 3," and where
this is a simpler form than was shown in the example above. In the
second case, the same functions of wavelength characteristics,
f(.lamda.).sub.e,ex and f(.lamda.).sub.e,em, each with values
particular to the module being normalized, are applied to the
predetermined normalization factor equation, using the
predetermined optimization parameters, to assign a normalization
factor for the module.
[0084] As such, for a given analyzer (or, more specifically, for
the module thereof), a set of normalization factors, s.sub.e, may
be determined, with each normalization factor corresponding to a
particular electrolyte sensor. Thus, after using the analyzer to
run a particular slide containing a sample thereon, the
normalization factor specific to the analyzer and the electrolyte
sensor contained on the slide may then be used, in conjunction with
the analyzer response to the electrolyte sensor and the slide's lot
calibration, to correct, or normalize the output value of the
analyzer into a concentration measurement of the electrolyte within
the sample.
Embodiment 2
[0085] In another embodiment, as discussed above, the overlap
between the spectra, e.g., the excitation and emission spectra, of
the module and the slide, e.g., the fluorophore or sensor, is
analyzed to determine the normalization factor for a particular
analyzer and electrolyte to be detected. The relative overlap
between the spectra is important because, for example, a particular
module may more effectively excite the fluorescence of an "early"
or "late" sensor (i.e., where there is greater overlap between the
spectra) or may less effectively excite the "late" or "early"
fluorescence of the sensor (i.e., where the spectra are less
overlapped) thereby introducing a bias (or biases) toward and/or
against a particular sensor's analyzer response. However, by
analyzing the overlap between the spectra of the module and the
"early" and "late" sensor spectra, such variations, or biases, can
be accounted for in determining the normalization factor for a
particular module and electrolyte (or the particular fluorophore
used to detect that electrolyte).
[0086] First, the module excitation spectrum data, relative
irradiance as a function of wavelength, f(.lamda..sub.i).sub.m,ex,
is collected as (wavelength, relative irradiance) ordered pairs and
is indexed accordingly. Similarly, the slide excitation spectrum
data, relative fluorescence as a function of excitation wavelength,
f(.lamda..sub.i).sub.e,early,ex and f(.lamda..sub.i).sub.e,late,ex,
is collected as (wavelength, fluorescence intensity) ordered pairs
in both early and late readings, respectively, as discussed above.
However, in this embodiment, the data need not be standardized or
mathematically normalized, although the data may be filtered. The
ratio of the overlaps of module and "early" and "late" slide
excitation spectra are calculated to obtain a function of
excitation spectra overlap:
f ( .lamda. ) m , e , l , c , ex , ovr = ( i = 1 N ex f ( .lamda. i
) m , ex .times. f ( .lamda. i ) e , l , c , late , ex i = 1 N ex f
( .lamda. i ) m , ex .times. f ( .lamda. i ) e , l , c , early , ex
) EQ 15 ##EQU00010##
In some embodiments, explicit consideration of slide lot/and sample
concentration c may be dropped.
[0087] Similarly, the module emission spectrum data, irradiance
sensitivity as a function of wavelength, f(.lamda..sub.i).sub.m,em,
is collected as (wavelength, relative irradiance sensitivity)
ordered pairs and is indexed accordingly and the slide emission
spectrum data, relative fluorescence as a function of emission
wavelength, f(.lamda..sub.i).sub.e,early,em and
f(.lamda..sub.i).sub.e,late,em, is collected as (wavelength,
fluorescence intensity) ordered pairs in both early and late
readings, respectively, as discussed above. As with the excitation
data, this data need not be standardized or mathematically
normalized, but may be filtered. The ratio of the overlaps of
module sensitivity and "early" and "late" slide emission spectra
are calculated to obtain a function of emission spectra
overlap:
f ( .lamda. ) m , e , l , c , em , ovr = ( i = 1 N em f ( .lamda. i
) m , em .times. f ( .lamda. i ) e , l , c , late , em i = 1 N em f
( .lamda. i ) m , em .times. f ( .lamda. i ) e , l , c , early , em
) EQ 16 ##EQU00011##
where, again, explicit consideration of slide lot l and sample
concentration c may be dropped in some embodiments.
[0088] Using the above data sets, "EQ 15" and "EQ 16," the ratios
of the overlaps of module and sensor excitation and emission
spectra can be used to fit a target normalization factor with an
analytical form similar to "EQ 14," where, dropping the module,
slide lot, and concentration indices, f(.lamda.).sub.e,ex,ovr (from
"EQ 15") is substituted for f(.lamda.).sub.e,ex and
f(.lamda.).sub.e,em,ovr (from "EQ 16") is substituted for
f(.lamda.).sub.e,em.
[0089] A simpler equation, the same as "EQ 4" but in different
form, is also considered:
s.sub.e=c.sub.e,1.times.f(.lamda.).sub.e,ex,ovr.times.f(.lamda.).sub.e,e-
m,ovr+c.sub.e,0 EQ 17:
as are more complicated forms involving linear and non-linear
combinations of the two spectrum overlap ratios and their product,
or any other suitable equations incorporating the principles of the
present disclosure.
[0090] As such, similarly to the previous embodiment, for a given
module, a set of normalization factors, s.sub.e, may be determined,
with each normalization factor corresponding to a particular
electrolyte sensor. Note again that the concentration subscript is
dropped if normalization factors are averaged for multiple
concentrations. Thus, after using the analyzer to run a particular
sample, the normalization factor specific to the analyzer and the
electrolyte sensor on the slide containing the sample may then be
used, in conjunction with the slide's lot calibration, to correct
or normalize the output value of the analyzer into a concentration
measurement of the electrolyte within the sample.
[0091] As can be appreciated, once the set of normalization
factors, s.sub.e, for a given module have been determined, e.g.,
according to any of the embodiments above, the particular module
need not be re-normalized by running a plurality of samples of
known concentration thereon. Rather, the sample need only be run on
the module to provide an analyzer response value, which, in
conjunction with the slide's lot calibration (determined
independently) and the normalization factor specific to that module
and electrolyte to be detected, is used to determine the
electrolyte concentration of the specific electrolyte within the
sample.
[0092] From the foregoing and with reference to the various figure
drawings, those skilled in the art will appreciate that certain
modifications can also be made to the present disclosure without
departing from the scope of the same. While several embodiments of
the disclosure have been shown and described, it is not intended
that the disclosure be limited thereto, as it is intended that the
disclosure be as broad in scope as the art will allow and that the
specification be read likewise. Therefore, the above description
should not be construed as limiting, but merely as exemplifications
of particular embodiments. Those skilled in the art will envision
other modifications within the scope and spirit of the claims
appended hereto.
* * * * *